Bayesian decision method of inspection and maintenance planning for deteriorating RC bridges

In order to determine the optimal inspection and maintenance plan for deteriorating reinforced concrete (RC) bridges and achieve the best balance between structural reliability level and life cycle cost, a Bayesian decision method based on risk-based inspection (RBI) was proposed. Based on the theoretical deterioration model of RC bridges in a chloride-ion erosion environment, a quantitative index system for corrosion damage assessment was developed. A stochastic deterioration process model and a reliability updating method based on a dynamic Bayesian network were established. By comprehensively considering the uncertainties of environmental parameters, damage detection, and maintenance decisions, the Bayesian network was expanded by introducing decision nodes describing inspection actions and utility nodes quantifying expected costs and benefits. A inspection and maintenance decision system based on a limited memory influence diagram (LIMID) was formed. The proposed method was applied to the inspection and maintenance planning of the RC bridge deck of an in-service composite girder bridge. Analysis results show that the theoretical corrosion failure probability of the bridge deck increases significantly with the extension of service life, reaching 19.4% and 45.5% in the 40th and 60th years, respectively. Necessary maintenance measures are urgently needed. By using the method in this article for inspection and maintenance planning, the optimal inspection times are obtained as the 18th, 33rd, 48th, and 61st years. The expected relative total cost is 260.3, including an inspection cost of 0.769, a maintenance cost of 187.8, and a failure cost of 71.7. Compared with the actual maintenance plan of the bridge, the cost is reduced by 36.3%. The relative total cost under the periodic inspection (PI) method is 271.1, and that under the reliability threshold (RT) method is 270.4. The proposed method provides an optimal solution.